Intro

I first identify rare types (<20 occurrences) and identify the types they optimally should be grouped with. I usually combined types within not across ecoregions.

## 
## RT1_Ireland and Northern Ireland                       RT1_Tundra 
##                               18                               19 
##                RT10_The Caucasus             RT12_Baltic province 
##                                3                                2 
##             RT2_Borealic uplands                 RT2_The Caucasus 
##                                3                               14 
##      RT3_Hellenic western Balkan                 RT3_The Caucasus 
##                                5                               11 
##      RT5_Hellenic western Balkan              RT6_Baltic province 
##                               13                                2 
##             RT6_Borealic uplands                     RT6_Pyrenees 
##                                4                                1 
##                 RT6_The Caucasus RT7_Ireland and Northern Ireland 
##                                6                                4 
##                     RT7_Pyrenees                 RT7_The Caucasus 
##                               19                               16 
## RT8_Ireland and Northern Ireland                       RT8_Tundra 
##                                7                               12 
##              RT9_Baltic province                     RT9_Pyrenees 
##                                4                                3 
##                        RT9_Taiga 
##                                2

I keep the rare RT1-based types, as very large rivers are quite distinct (e.g. Borgwardt et al. 2019).
In the following I adress the rare types ordered by ecoregions.

Baltic province

It is very surprising to find rivers of the type RT12 Mediterranean temporary and very small so far in the north. I assume that this is an erroneous assignment. As we are only dealing with two segments, I remove them from the data. The rivers of types RT6 and RT9 are combined with those of the same geology and altitude but different size (i.e. RT7 and RT8)

The Caucasus

Only a few stream segments at the very edge of our data fall within this ecoregion. I remove it from the data set.

Borealic uplands

Both rare types (RT2 and 6) are calcareous medium-large types. It would be a possibility to combine them into one. Given the low prevalence of each type though (3 and 4 instances), the combined type would still be rare. Hence I rather combine them with the respective small types.

Pyrenees

As with the Baltic province, RT6 and RT9 are combined with RT7 and RT8.

Taiga

The only instance of the RT9 type (mid-altitude siliceous very small small) drains in to RT5 (lowland siliceous very small small). We will hence consider them as one type.

Tundra

The area covered by BRT and the Illies region “Tundra” is small. I add them to the borealic Uplands

Ireland and Northern Ireland

Here we find the opposite situation to the Pyrenees and the Baltic province: RT7/8 are rare and RT 6/9 more common. The approach, however, remains the same. We combine the mid-altitude streams to one type per bedrock, irrespective of size.

Hellenic western Balkan

In this ecoregion small lowland rivers are rare because they are mostly temporary and thus fall under RT12. I will combine RT3 and RT5 to Hellenic very small - small perennial lowland rivers.

System

## R version 4.1.0 (2021-05-18)
## Platform: x86_64-w64-mingw32/x64 (64-bit)
## Running under: Windows 10 x64 (build 19042)
## 
## Matrix products: default
## 
## locale:
## [1] LC_COLLATE=German_Germany.1252  LC_CTYPE=German_Germany.1252   
## [3] LC_MONETARY=German_Germany.1252 LC_NUMERIC=C                   
## [5] LC_TIME=German_Germany.1252    
## 
## attached base packages:
## [1] stats     graphics  grDevices utils     datasets  methods   base     
## 
## other attached packages:
## [1] stringr_1.4.0  tmap_3.3-2     dplyr_1.0.7    magrittr_2.0.1 sf_1.0-1      
## 
## loaded via a namespace (and not attached):
##  [1] tidyselect_1.1.1        xfun_0.24               bslib_0.2.5.1          
##  [4] purrr_0.3.4             lattice_0.20-44         leaflet.providers_1.9.0
##  [7] vctrs_0.3.8             generics_0.1.0          viridisLite_0.4.0      
## [10] htmltools_0.5.1.1       stars_0.5-3             s2_1.0.6               
## [13] base64enc_0.1-3         yaml_2.2.1              utf8_1.2.1             
## [16] XML_3.99-0.6            rlang_0.4.11            e1071_1.7-7            
## [19] jquerylib_0.1.4         pillar_1.6.1            glue_1.4.2             
## [22] DBI_1.1.1               sp_1.4-5                RColorBrewer_1.1-2     
## [25] wk_0.5.0                lifecycle_1.0.0         raster_3.4-13          
## [28] htmlwidgets_1.5.3       codetools_0.2-18        leafsync_0.1.0         
## [31] evaluate_0.14           knitr_1.33              crosstalk_1.1.1        
## [34] parallel_4.1.0          class_7.3-19            fansi_0.5.0            
## [37] leafem_0.1.6            Rcpp_1.0.7              KernSmooth_2.23-20     
## [40] classInt_0.4-3          lwgeom_0.2-6            leaflet_2.0.4.1        
## [43] jsonlite_1.7.2          abind_1.4-5             png_0.1-7              
## [46] digest_0.6.27           stringi_1.7.2           tmaptools_3.1-1        
## [49] grid_4.1.0              tools_4.1.0             sass_0.4.0             
## [52] proxy_0.4-26            tibble_3.1.2            dichromat_2.0-0        
## [55] pacman_0.5.1            crayon_1.4.1            pkgconfig_2.0.3        
## [58] ellipsis_0.3.2          assertthat_0.2.1        rmarkdown_2.9          
## [61] R6_2.5.0                units_0.7-2             compiler_4.1.0

References

Borgwardt, Florian, Patrick Leitner, Wolfram Graf, and Sebastian Birk. 2019. “Ex Uno Plures–Defining Different Types of Very Large Rivers in Europe to Foster Solid Aquatic Bio-Assessment.” Ecological Indicators 107: 105599.